questions

---

Latest Posts

OPUpdated

Is onFinish called when stream is aborted?

How to implement message interruptions with stop()?

How does human in the loop work?

Does finishReason "stop" always send for streamText?

How to get UIMessagePart creation timestamp with streamText?

How are AI Gateway model strings converted to language models?

Can I send system role messages mid-conversation?

Is there a notion of tool context?

How do I implement tool call streaming?

How do I create agents with the AI SDK?

How to create chat bubble UI with version 6?

Why don't two useChat hooks with same ID share state?

Simplest Next.js setup for message streaming?

Why is "No output generated" error thrown?

How does response caching work?

How does the SDK handle errors?

How are tokens counted and tracked?

How to set up prompt caching for multiple AI models?

How to configure reasoningEfforts in my agent loop?